One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection

Memory technologies for in-memory computing (IMC) based neural network (NN) operations in edge devices face two primary challenges. First, these technologies often exhibit an excessive average power due to the high currents in the memory’s low and high resistance state (LRS and HRS). Seco...

Full description

Saved in:
Bibliographic Details
Main Authors: Shreyas Deshmukh, Ankit Bende, Diti Sanghai, Vivek Saraswat, Anmol Biswas, Abhishek Kadam, Shubham Patil, Ajay Kumar Singh, Veeresh Deshpande, Udayan Ganguly
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of the Electron Devices Society
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10771781/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849412517237882880
author Shreyas Deshmukh
Ankit Bende
Diti Sanghai
Vivek Saraswat
Anmol Biswas
Abhishek Kadam
Shubham Patil
Ajay Kumar Singh
Veeresh Deshpande
Udayan Ganguly
author_facet Shreyas Deshmukh
Ankit Bende
Diti Sanghai
Vivek Saraswat
Anmol Biswas
Abhishek Kadam
Shubham Patil
Ajay Kumar Singh
Veeresh Deshpande
Udayan Ganguly
author_sort Shreyas Deshmukh
collection DOAJ
description Memory technologies for in-memory computing (IMC) based neural network (NN) operations in edge devices face two primary challenges. First, these technologies often exhibit an excessive average power due to the high currents in the memory&#x2019;s low and high resistance state (LRS and HRS). Second, their vulnerability to tampering compromises device reliability in security-sensitive Internet of Things (IoT) applications. To tackle these challenges, we propose the use of one-time programmable memory (OTPM) for IMC-based thermal attack-secured and ultra-low powered NN accelerator with write disablement to prevent re-programming/tampering. In this article, first, we show that HRS for OTP memory is exceptionally high (~190 M<inline-formula> <tex-math notation="LaTeX">$\Omega $ </tex-math></inline-formula>), but the higher current in LRS (~100 mA) presents a challenge of high average power. We address this challenge by introducing a select transistor gate voltage biasing scheme with the 1T-1OTPM bit cell to reduce LRS current significantly while still trading off the excessive HRS/LRS ratio of 1.5 M to about 1 k. This reduces average power by <inline-formula> <tex-math notation="LaTeX">$\sim 1000\times $ </tex-math></inline-formula> without degrading CIFAR-10 classification performance. Second, we evaluate the impact of OTPM size scaling on HRS and LRS distributions. Though scaling reduces HRS current, we show that moderately scaled 10F<inline-formula> <tex-math notation="LaTeX">$\times 10$ </tex-math></inline-formula>F OTPM bit cells produce the lowest maximum current in HRS compared to the more aggressively scaled 1F<inline-formula> <tex-math notation="LaTeX">$\times 1$ </tex-math></inline-formula>F, which suffers from significant process-induced variability. Third, we show the robustness to temperature, which enables security from thermal attacks.
format Article
id doaj-art-b1179f8afcd149c7a234e2bd1ff32a07
institution Kabale University
issn 2168-6734
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of the Electron Devices Society
spelling doaj-art-b1179f8afcd149c7a234e2bd1ff32a072025-08-20T03:34:25ZengIEEEIEEE Journal of the Electron Devices Society2168-67342025-01-011385486210.1109/JEDS.2024.350875910771781One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault InjectionShreyas Deshmukh0https://orcid.org/0000-0003-0086-1276Ankit Bende1Diti Sanghai2Vivek Saraswat3https://orcid.org/0000-0001-9191-1632Anmol Biswas4https://orcid.org/0000-0002-2363-5864Abhishek Kadam5https://orcid.org/0000-0001-6106-5623Shubham Patil6https://orcid.org/0000-0003-2375-481XAjay Kumar Singh7https://orcid.org/0000-0002-6170-1340Veeresh Deshpande8Udayan Ganguly9https://orcid.org/0000-0002-1498-5993Department of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, IndiaMemory technologies for in-memory computing (IMC) based neural network (NN) operations in edge devices face two primary challenges. First, these technologies often exhibit an excessive average power due to the high currents in the memory&#x2019;s low and high resistance state (LRS and HRS). Second, their vulnerability to tampering compromises device reliability in security-sensitive Internet of Things (IoT) applications. To tackle these challenges, we propose the use of one-time programmable memory (OTPM) for IMC-based thermal attack-secured and ultra-low powered NN accelerator with write disablement to prevent re-programming/tampering. In this article, first, we show that HRS for OTP memory is exceptionally high (~190 M<inline-formula> <tex-math notation="LaTeX">$\Omega $ </tex-math></inline-formula>), but the higher current in LRS (~100 mA) presents a challenge of high average power. We address this challenge by introducing a select transistor gate voltage biasing scheme with the 1T-1OTPM bit cell to reduce LRS current significantly while still trading off the excessive HRS/LRS ratio of 1.5 M to about 1 k. This reduces average power by <inline-formula> <tex-math notation="LaTeX">$\sim 1000\times $ </tex-math></inline-formula> without degrading CIFAR-10 classification performance. Second, we evaluate the impact of OTPM size scaling on HRS and LRS distributions. Though scaling reduces HRS current, we show that moderately scaled 10F<inline-formula> <tex-math notation="LaTeX">$\times 10$ </tex-math></inline-formula>F OTPM bit cells produce the lowest maximum current in HRS compared to the more aggressively scaled 1F<inline-formula> <tex-math notation="LaTeX">$\times 1$ </tex-math></inline-formula>F, which suffers from significant process-induced variability. Third, we show the robustness to temperature, which enables security from thermal attacks.https://ieeexplore.ieee.org/document/10771781/Artificial neural networkin-memory computationnon-volatile memoryone time programmable memorysecured memory technology
spellingShingle Shreyas Deshmukh
Ankit Bende
Diti Sanghai
Vivek Saraswat
Anmol Biswas
Abhishek Kadam
Shubham Patil
Ajay Kumar Singh
Veeresh Deshpande
Udayan Ganguly
One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection
IEEE Journal of the Electron Devices Society
Artificial neural network
in-memory computation
non-volatile memory
one time programmable memory
secured memory technology
title One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection
title_full One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection
title_fullStr One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection
title_full_unstemmed One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection
title_short One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection
title_sort one time programmable memory for ultra low power ann inference accelerator with security against thermal fault injection
topic Artificial neural network
in-memory computation
non-volatile memory
one time programmable memory
secured memory technology
url https://ieeexplore.ieee.org/document/10771781/
work_keys_str_mv AT shreyasdeshmukh onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT ankitbende onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT ditisanghai onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT viveksaraswat onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT anmolbiswas onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT abhishekkadam onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT shubhampatil onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT ajaykumarsingh onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT veereshdeshpande onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection
AT udayanganguly onetimeprogrammablememoryforultralowpoweranninferenceacceleratorwithsecurityagainstthermalfaultinjection